People spend millions on features they don’t need
Salesforce ($CRM) charges you to enter the building, then again to turn on the light. I call that greed.
Traditional software does not sell the job. It sells the bundle. You need a pipeline. The invoice includes forecasting, a service desk, a marketing cloud, a sandbox, storage, and an AI add-on for the person who logs in twice a month. You pay for the menu. The team uses three items.
Salesforce is the clean version of that, because the meter is public.
In June 2025 it raised prices on the core clouds by about 6%, and said the increase was to pay for AI. The AI was not thrown in. Agentforce add-ons start at $125 per user per month, on top of Enterprise or Unlimited. The full Agentforce 1 edition starts at $550 per user per month. If you would rather be metered, Flex Credits run 20 credits, or $0.10, every time an agent updates a record, runs a workflow, or resolves a case. A pack of 100,000 credits is $500. By early 2026 there were three pricing models on the same product. In April 2026 Truist still called Agentforce a pricing problem.
That is the nickel and dime. A seat. Then a higher edition so the seat can do the thing you thought you bought. Then an add-on. Then a charge per action. The renewal is not a vote that the tool got better. It is the cost of leaving.
I am not betting that companies stop keeping records. I am betting that a lot of the Salesforce bill is features and meters they do not need, and that a smaller system takes the seat.
For a company sitting on it, I build the CRM they actually use. $5,000 to $10,000, plus maintenance. Then I organize the pipeline. The stages, the handoffs, the follow-up. Not a $125 add-on so a model can update a field you already had.
Same view in the stock. I bought way out of the money puts on $CRM expiring next year. They pay if the bundle stops winning. They expire worthless if customers keep paying for the lights.
You can check my record. On January 27, 2025 I wrote on LinkedIn, “SAAS will die.” On October 14, 2026, three weeks ago, I wrote, “I disagree with that now.” On November 2 I wrote, “SAAS companies no longer have the edge they used to.” So I said it, took it back, and came partway back. Nothing died. From that January post to the November 4 close, Salesforce is down 34.6%, the IGV software ETF is up 1.4%, and the S&P 500 is up 27.9%. This bet is narrower than what I said in 2025.
Source: LinkedIn, January 27, 2025 · LinkedIn, October 14, 2026 · X, November 2, 2026 · Yahoo Finance CRM data · IGV data · S&P 500 data
Same with AI. On October 7 I wrote, “It’s a big bubble.” On October 13 and 15 I posted puts on CoreWeave, Nvidia, Oracle, Meta, AMD, Broadcom and Palantir, with the strikes, all expiring September 17, 2027. From the close the day I named each one to the November 4 close, all seven fell: CoreWeave 37.8%, Meta 17.4%, Palantir 14.4%, Oracle 13.6%, AMD 13.1%, Nvidia 12.4%, Broadcom 5.5%. The S&P 500 fell 1.5% from October 13. Those are stock moves, not my result on the puts. I have since taken profit on some and bought others further out of the money. All seven are still above the strikes I posted, and the S&P 500 is still up 12.3% this year, so this call isn’t settled.
Source: LinkedIn comment, October 7, 2026 · Puts, October 13, 2026 · Puts, October 15, 2026 · Yahoo Finance data: CRWV · META · PLTR · ORCL · AMD · NVDA · AVGO · My October 31 article
Some of it I got wrong or can’t prove. In April 2025 I wrote, “AI can take 99% of white collar jobs.” I don’t believe that now. In March 2025 I wrote, “Bad news for NVIDIA again.” Nvidia is up 75.5% since. And on October 16 I wrote, “People have overestimated the demand for AI and data centers.” The reporting points to payoff and financing worries instead, and Nvidia’s CEO called that demand very strong.
Source: LinkedIn, April 22, 2025 · LinkedIn, March 3, 2025 · LinkedIn, October 16, 2026 · CNBC, October 30, 2026
Sources: Salesforce, June 17, 2025, Agentforce add-ons from $125 per user per month and Agentforce 1 Editions from $550. Salesforce Ben, May 15, 2025, Flex Credits at $0.10 per action and $500 per 100,000 credits. IT Brew, CX Today, and CIO, June 2025, on the 6% core-cloud increase. SaaStr, February 17, 2026, on three Agentforce pricing models. Yahoo Finance, April 27, 2026, Truist on the Agentforce pricing problem.
Disclosure: This is my opinion, not investment advice. I don’t know your situation. As of November 5, 2026, I hold way out of the money put options on Salesforce ($CRM) expiring next year. Those puts make money if the stock falls. I may sell them at any time without saying so. I hold put options on AI stocks as well, which make money if those stocks fall. Since the October posts named above I have taken profit on some and bought some further out of the money. I may sell or change these at any time without saying so. I also sell a custom CRM build-out, priced at $5,000 to $10,000 plus maintenance, and I organize the pipeline. I get paid if you hire me.